A guide to building longrange wireless links ( over 1km ) using standard hardware and remaining within legal power limits. This is information that some manufacturers have known all along, but that I've only recently understood.
I recently managed a successful 5km link across town, using just standard access points and antenna. While working out all of the signal loss calculations to see if the link was even possible, I started to understand how long range links work, and how manufacturers such as Cisco have been able to offer long range links while staying within the legal transmission power limits for Wi-Fi.
I had always assumed that the maximum range a wireless connection could be was about 1km (based on uk power limit) and this was in line with the practical range tests I had carried out. To get increased range, always required sticking a big antenna on the Access Point, pushing the overall signal well over the uk limit (20dbm).
While carrying out the 5km link test I found that the snr (signal to noise ratio) was not the really critical factor, but that it was the received power level that decided if a link was possible.
The power usually expressed in mW or db that the wireless device transmits at.
The level of signal the wireless device needs to receive in order to make a connection. Most wireless devices require a stronger signal to pick up a connection at 11Mbps than one at 1Mbps.
Antenna don't increase the transmission power, but focus the signal more. So instead of transmitting in every direction (including the sky and ground) antenna focus the signal usually either more horizontally or in one particular direction. This gain is measured in decibels.
This is the strength of the background noise that the wireless device is picking up.
Signal to Noise Ratio
The difference between the strength of the background noise and the strength of the signal. If the noise is greater than the signal, then no signal will be picked up. If the noise is only just less than the signal then there will be lots of lost packets. Ideally at least 5db difference between the noise and signal will allow for a clear connection.
EU Power Limit
In Europe the maximum signal that your antenna can transmit in ANY direction is 20 decibels. This is calculated by adding the wireless device transmit power(in decibels) to the antenna gain (in decibels).
The basic premise for long range links is
Transmit Power is independant of Receive Sensitivity
I had always assumed (wrongly) that an access point that transmitted at a higher power, would receive the signal at a higher power as well. This is not the case. Increasing the power that an access point (or any wireless device) transmits at, does not affect the level of signal it needs to receive in order to make a connection.
When you stick an antenna onto a wireless device, the antenna gain acts on both the transmit power and receive sensitivity, so you are not only sending your signal further (in a more focused direction), but your are also able to listen to weaker signals. The EU regulations only limit the transmit signal strength, not the receive sensitivity, so we need a way of increasing the receive sensitivity, but not increasing the transmit signal beyond the 20db limit.
Most access points and wireless devices in Europe transmit at 15 - 17.5 db. (Some may claim 20db , but this includes the gain of the small antenna they are using) This would limit the antenna gain to just 2.5 - 5db (or less), which is not going to affect the range significantly. In free space (ie outside, with clear line of sight ) for every 6db of increased signal strength or receive sensitivity you can double the range between the devices.
Some wireless devices however, let you reduce the transmit power, which allows us to use a more powerful antenna. This does not increase the overall strength of the signal , but does increase the receive sensitivity. If we do this at both ends of the link, we can increase the range, while staying within legal limits.
The Maths Explained
In order to get a signal from point A to point B, the signal arriving at the wireless device B needs to be greater than the receive sensitivity of that device
TPowerA + AntGainA - SignalLossAtoB + AntGainB > ReceiveSensB
TPowerA = Transmit Power of A
AntGainA = Antenna Gain of A
AntGainB = Antenna Gain of B
SignalLossAtoB = Signal Loss between A and B
ReceiveSensB = Receive sensitive of B
In order to see how far we can get a signal, we need to calculate how much signal loss we can cope with, and still get a strong enough signal to B.
Maximum SignalLossAtoB = TPowerA + AntGainA + AntGainB - ReceiveSensB
As an example, if we base our figures on a 15db wireless device, 5db antenna at each end and -92 db receive sensitivity we get.
Maximum SignalLossAtoB = 15 + 5 + 5 - -92
Maximum SignalLossAtoB = 117db
In practice 117db should give you a range of 1-2 km (In theory it should be over 5km). If we now look at reducing the transmit power to 7db, and using 13db antenna on each end thus keeping the transmit signal strength to 20db, we get
Maximum SignalLossAtoB = 7 + 13 + 13 - -92
Maximum SignalLossAtoB = 125db
This gives us over twice the range of before, yet still keeps the signal strength legal.
Once I had worked out how to make long range links, I started to look round for wireless kit that would support long range point to point links. I was very disappointed to find not only the market quite thin, but the prices very high for such kit. Update The latest firmware for the D-Link DWL900AP+ Access point supports variable transmit power, making it the cheapest entry in the long range link arena by far. I remembered that the old v1.1 Wap11 from linksys could have its power level adjusted using a 3rd party Amtel utility, but on looking at the spec for the old Wap11 v1.1, the receive sensitivity is so poor, that it is not suitable for long range links.
Kit Pros Cons
Cisco Aironet 350 Market Leader, Proved reliablity, Maximum Claimed Range 40.2km Oh so Expensive
3Com Wireless Bridge Maximum Claimed Range 16.9km Expensive, less range than Cisco
D-Link DWL-900AP+ Cheap and well supported Newcomer to the longrange arena
 Device power is actually measured in dbm and antenna gain in dbi, but if you already know this, then this article is probably not for you.
 Freespace loss is also dependent on an unobstructed fresnal zone, but I don't really want to go into that here.
 In order to simplify the maths I'm ignoring cable loss between the wireless device and the antenna, as this can be simply treated as reduced antenna gain..
© 2002-2006 wireless.gumph.org (ppl who wrote this i have jsut posted it here 2 help others